Camelia, the Perl 6 bug

IRC log for #bioclipse, 2010-03-03

| Channels | #bioclipse index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
04:56 egonw joined #bioclipse
07:52 sneumann joined #bioclipse
08:02 Gpox joined #bioclipse
08:12 masak joined #bioclipse
08:52 annzi joined #bioclipse
08:52 olass joined #bioclipse
09:04 jonalv joined #bioclipse
09:22 egonw joined #bioclipse
09:23 egonw moi
09:23 egonw n
09:27 masak \
09:27 masak o
09:36 jonalv oh
09:36 jonalv hai
09:43 jonalv egonw: I have been thinkging regarding your extension point
09:43 jonalv egonw: or should I say our
09:43 egonw I was hoping for 'your' :)
09:43 jonalv egonw: I am not 100 % sure I like it.
09:44 jonalv it will create a need for either chemoinformatics and bioinformatics +some others to depend on bioclipse.rdf
09:45 jonalv I regarded bioclipse.rdf as something that could extend bioclipse...
09:45 egonw ah, no, it does not
09:45 egonw we are not going to make the cheminfo or bioinfo feature contain the extension
09:45 jonalv or it will create features like bc.chemoinformatics.rdf and bc.bioinformatics.rdf
09:45 egonw +s
09:46 egonw those will go into separate extensions: RDF-Cheminformatics Integration Feature
09:46 jonalv egonw: which would be a feature explosion...
09:46 egonw and
09:46 egonw RDF-Bioinformatics Integration Feature
09:46 egonw yes, it's one way or the other
09:46 egonw well, not that I think the features will really explode
09:47 jonalv is it impossible to see which Strings are potential IBioObjects?
09:47 egonw but there will be a few more
09:47 egonw what Strings?
09:47 egonw it takes graphs
09:47 egonw RDF graphs
09:47 jonalv egonw: When I doubleclick a box in the editor I am not double clicking a graph am I?
09:47 jonalv I am doubleclicking a node
09:48 jonalv Isn\t the smiles a property of that node?
09:48 egonw no
09:48 egonw it's connected via a edge
09:48 jonalv egonw: what is it then?
09:48 egonw aka, triple
09:48 egonw aka, jena statement
09:48 egonw moreover, it may need not the SMILES but the InChI
09:48 jonalv so maybe I should doubleklick the smiles instead?
09:49 jonalv egonw: but is the smiles visualised in the editor?
09:49 egonw in general, it will need all triples where "zest:clicked ?p ?o"
09:49 egonw double click the SMILES literal will make you loose all other interesting props
09:50 jonalv egonw: my point is that if we could reuse the functionality for converting any given string to a bioobject then we gain alot
09:51 egonw yes, but I don't feel much for solving two distinct problems with the same solution, because just of code reuse
09:51 egonw the problems *are* different
09:52 jonalv egonw: it's not only code resue but also to avoid feature explosion. But it they _are_ different then that is not the way to do it...
09:52 egonw I do not feel much for introducing a whole range of RDF-cheminformatics and RDF-bioinformatics file formats
09:52 jonalv egonw: I am just having problems seeing that they are different
09:53 jonalv egonw: and I am not sure I want to create a new feature for each bioclispe feature containing an bioobject. That is a lot of features...
09:54 egonw in your suggested approach you need to write format detection for new formats, write readers for those, etc
09:54 jonalv egonw: hang on we need to parse that resource in both solutions, right?
09:55 egonw I don't see why you need to "create a new feature for each bioclispe feature containing an bioobject" ??
09:55 egonw jonalv: no, not is my suggested solution
09:56 jonalv egonw: if you want to be able to open the corresponding rdf entity fror each of those bioobjects from the rdf editor then you are going to need a feature for each of them that contains the rdf specific code for that
09:56 egonw no, not even that
09:57 jonalv egonw: not a feature for each object but a feature for each feature containing objects
09:57 egonw you only need a feature matching the features those IBioObjects are in
09:57 egonw right
09:57 jonalv that's what I tried to say...
09:57 egonw so, we now have cheminfo, bioinfo, and speclipse
09:57 egonw so three integration features...
09:57 egonw perhaps a small fire?
09:58 jonalv well, I don't like the genereal solution. Any day now someone will addd a similar feature and then you have to create 2 new features for each feature you create
09:58 jonalv it scales m*n I don't like that...
09:58 egonw 2?
09:58 egonw why 2??
09:59 jonalv egonw: if I want to add a feature with bioobjects after this I am going to need to create my-feature and my-feature.rdf
09:59 jonalv I am just seeing me gonna have to create my-feature.xyz
09:59 jonalv some other day
09:59 jonalv I don't like this genreal trend
09:59 jonalv masak: you can see my point right?
09:59 egonw the rdf integration is optional, not required
09:59 jonalv egonw: I don't think that is an excuse
10:00 egonw jonalv: I think I see your point clearly, but that does not mean I have to agree
10:00 jonalv egonw: hm is that possible... Can one see this point and not agree? :)
10:01 egonw jonalv: so, are you in or out?
10:01 masak jonalv: sorry, not really following along in the conversation right now.
10:01 jonalv egonw: well can you make one more short attempt at convincing me that these are two different problems?
10:02 egonw nope, I could only reiterate my arguments
10:02 jonalv egonw: All I am seing right now is that you locate the SMILES string and then open that. Which is not different from the other case...
10:03 jonalv unless the locate part is the tricky part...
10:03 egonw please do look beyond that...
10:03 egonw again, it's *not* a string being passed, but a RDF graph
10:03 egonw also, the locate part will hard code the dependency on cheminfo, bioinfo, speclipse
10:04 jonalv egonw: yes so you keep saying. Does this mean that we might be intrested in more than one of the nodes in that RDF graph?
10:04 egonw it's not hard, but domain specific, and cannot be implemented in the extension XML
10:04 egonw jonalv: yes
10:04 jonalv egonw: hm so we might want to open more than just say a smiles literal...
10:05 jonalv egonw: I can't see an example for that but if that is the case then "I can see your point...
10:05 egonw indeed.... I did mention that we pass a graph, right?
10:05 egonw here's the use case:
10:05 egonw (takes some secs to get the proper URI)
10:06 jonalv egonw: yes, but in your example you are only intrested in one node in that graph. So why not only pass that node and be done with it I ahve been thinking...
10:06 egonw lesson learned: giving simple examples only backfires
10:06 egonw (I should make a nice haiku for that)
10:07 jonalv hehe
10:07 egonw http://saml.rilspace.com/content/​nmr-sparql-search-problem-solved
10:07 zarah egonw's link is also http://tinyurl.com/y8hlr9a
10:07 egonw note the nmr:hasPeak
10:08 egonw a spectrum has many peaks, each of which hasShift and hasIntensity
10:08 jonalv egonw: ahh
10:09 jonalv egonw: right so to handle this as a String we would have to first create a string of it and then parse that string and open that
10:09 jonalv hm not a super duper solution I agree
10:09 egonw so, we're actually talking here about a node "zest:selected rdf:type foo:Spectrum", where what we want to extract is 1+ triples "zest:selected ?p ?o", but even 'properties of ?o too
10:09 jonalv string being something like cml or so
10:09 egonw no
10:09 egonw can't do that
10:10 egonw that would make the RDF editor depend on all domain feature
10:10 egonw s
10:10 jonalv egonw: yes I guess it would...
10:10 jonalv hm, this is a mess
10:10 jonalv but you have convinced me
10:10 egonw only if you are afraid of more features
10:11 egonw ok, good, then I am heading now for coffee
10:11 jonalv k
10:11 egonw and refilling our pellet burner
10:11 egonw will be back later
10:11 jonalv cya
10:34 jonalv zarah: ask egonw Why does the JenaRDFLiteral not have the value of the actual literal?
10:34 zarah Consider it noted.
11:40 egonw_ joined #bioclipse
11:41 egonw_ jonalv: ping
11:52 egonw @tell jonalv is it common to show the object itself in the Props View too, instead of just its properties? I can add it, if you want it...
11:52 zarah Consider it noted.
11:52 zarah egonw: You have new messages. Write '/msg zarah @messages' to read them.
12:07 edrin joined #bioclipse
12:09 edrin greetings
12:12 egonw hi edrin
12:13 egonw annzi: have you had reply from Maris on the SPARQL output you sent him?
12:19 samuell joined #bioclipse
12:21 annzi egonw: no
12:21 egonw ack, will ping him
12:27 egonw annzi: see my email
12:28 egonw the sequences are not returned
12:28 egonw or not stored in the output, not sure
12:28 egonw and two other comments
12:28 egonw but the sequence is important for Maris to proceed
12:45 egonw olass: are your Bio-IT slides online?
12:46 olass nope
12:46 olass but I could put them up
12:46 olass after lunch :)
12:46 * olass is hungry
12:46 egonw please do
12:46 egonw then I can point people to it :)
12:57 jonalv zarah: just tell me
12:57 zarah jonalv: 1 h 5 m 22 s ago, egonw said is it common to show the object itself in the Props View too, instead of just its properties? I can add it, if you want it...
12:58 jonalv egonw: I would like to show the property value or something yea
12:58 egonw even if they are visible in the editor too?
12:58 egonw I'm on it...
12:58 jonalv egonw: yea
12:58 egonw hang on
13:01 masak joined #bioclipse
13:02 samuell all: I added some fields that users registering to the bioclipse.net website (for forums etc) can enter if they want. See http://www.bioclipse.net/user/register
13:02 zarah samuell's link is also http://tinyurl.com/ygdymfp
13:02 samuell Any comments?
13:03 samuell (Thinking it will be motivating for users to engage and help out in the forums etc. if they can at least have a link to their site, at their personal page)
13:03 samuell etc.
13:04 edrin samuell, i am usually reading the mails you wrote to the mailing list, thanks for all the work regarding the wiki and webpage.
13:04 edrin regarding the forum on the bioclipse.net page, can you link or sync it to the sf mailing list threads?
13:05 edrin so one can respond to posts via email and vice versa?
13:05 samuell edrin: Hmm ... thanks for the idea
13:06 samuell I'd have to check it up if that's possible
13:06 edrin it's just a suggestion to not have to places to discuss ...
13:06 samuell Yep ...
13:06 edrin two" even
13:07 samuell The basic idea though, has been to try to separate developers and users discussions a bit ... but I see your point, as there is now a developers forum on the web ...
13:07 samuell It's a good point, indeed
13:08 samuell (With the above, I mean, that the thought was to gather users discussions/support on the website, since many users might not easily find their way to the mailing lists)
13:08 edrin yes sure :)
13:09 samuell We have to think through this. I'll check up what options are there in Drupal for syncing mailing lists.
13:09 samuell syncing _with_ mailing lists.
13:11 jonalv egonw: so the selecting a literal does not give a literal in the selection event. You are aware of this?
13:11 egonw it's going to give the value and possible properties like Lang and Data Type
13:12 jonalv egonw: okey you are working on that now?
13:12 egonw yeah...
13:12 egonw commit pending...
13:12 jonalv egonw: cool
13:12 egonw eclipse was a bit slow booting...
13:12 egonw and got distracted by updating the matching paper latex source code :)
13:13 jonalv egonw: okey
13:17 jonalv Gpox: where in Eclipse was the command for removing triling white space and converting tbas to spaces?
13:18 CIA-57 bioclipse.rdf: Egon Willighagen master * rf41c226 / (2 files): Added value field, so that a literal value can be shown in the Props View too - http://bit.ly/9g9s4S
13:18 Gpox If you go to properties search for keys and then Remove Trailing Whitespace
13:19 jonalv Gpox: oh so it wasn't in a menu alternative?
13:19 Gpox no
13:19 Gpox you have to bind it to a key combination
13:20 samuell all: I made the dev forum on the website invisible now until we've figured this out. (Thanks edrin for the remark!)
13:20 jonalv Gpox: right how do I find an empty key combinatiopn?
13:22 Gpox good question, I have it on shift+command+space
13:30 jonalv egonw: so the selection of a literal still just returns an integer...
13:30 jonalv egonw: or _the_ literal I mean
13:31 egonw still have to check how literals are sent around...
13:31 egonw jonalv: that patch is pending...
13:31 jonalv egonw: sounds good
13:31 * egonw is bad at doing two things at the same time...
13:31 egonw want to send around the draft for the LODD paper
13:32 olass egonw: I need to talk to you about getting the stuff into a QSAR project tomorrow
13:32 olass need some pointers for how to start
13:32 olass let's discuss IRL tomorrow after annzis presentation
13:32 olass if OK with you?
13:32 olass and annzi?
13:32 egonw yes
13:33 egonw annzi: would make two nice slides... discuss the SPARQL queries for Martin and Maris...
13:35 annzi olass: fine with me
13:35 olass gut gut
13:35 olass supergut
13:35 olass alls klar
13:36 olass na ja
13:38 edrin samuell: maybe this helps: http://drupal.org/node/40777 (didn't read it in detail, though)
13:38 samuell edrin: Thanks, yep, something like this would be the solution. Will have to look into it.
13:55 CIA-57 bioclipse.rdf: Egon Willighagen master * r59e6ecb / (2 files): Use Literal instead of String for literal nodes in the RDF graph - http://bit.ly/bQlfQz
13:55 CIA-57 bioclipse.rdf: Egon Willighagen master * r1070721 / plugins/net.bioclipse.rdf.ui/src/net/b​ioclipse/rdf/ui/editors/RDFEditor.java : Send around a IRDFLiteral selection event when literals are selected - http://bit.ly/csYUMx
13:55 egonw jonalv: there you go
13:55 egonw sorry for the delay
14:04 jonalv egonw: pull request about properties sent
14:06 egonw received
14:08 jonalv egonw: regarding line 171-173 in the RDFEditor class. You know better than that right?
14:12 egonw via activator?
14:12 egonw I just keep forgetting if I should use the Java or the JS version
14:13 jonalv egonw: well put it like this, if you are writing Java use the Java version. If youare gonna publish it into the JavaScript environment use the JavaScript one
14:13 egonw ok, clear enough
14:15 jonalv egonw: your mail seems to lack an attachement... :(
14:15 egonw oh no, not again...
14:15 egonw gmail--
14:16 jonalv egonw: do you forget to add it or what are you doing wrong?
14:16 egonw no, I'm 100% positive I added it...
14:16 egonw it just doesn't attach them :(
14:16 jonalv sounds strange
14:18 egonw I particularly find it bloody annoying
14:23 egonw joined #bioclipse
14:25 jonalv egonw: seems chrome was the problem
14:25 egonw indeed
14:26 egonw chrome--
14:26 masak chrome--
14:35 egonw jonalv: thanx, one down!
14:35 CIA-57 bioclipse.rdf: jonalv master * rfc3a069 / plugins/net.bioclipse.rdf.ui/META-INF/MANIFEST.MF :
14:35 CIA-57 bioclipse.rdf: Removed version number.
14:35 CIA-57 bioclipse.rdf: We usually don't specify version number + I had an older version so it didn't
14:35 CIA-57 bioclipse.rdf: work without removing it foe me. - http://bit.ly/cOQYXR
14:35 CIA-57 bioclipse.rdf: jonalv master * r9bf8617 / (3 files in 2 dirs): Added PropertySources to RDF domain objects - http://bit.ly/cQBiTq
14:36 samuell joined #bioclipse
14:40 egonw jonalv: why doesn't it show the value? did you decide against that after all?
14:40 jonalv egonw: oh right, uhm will fix in next version
14:58 jonalv egonw: okey fixed
15:05 CIA-57 bioclipse.rdf: jonalv master * rcdf221e / plugins/net.bioclipse.rdf/src/net/bio​clipse/rdf/model/JenaRDFLiteral.java : added value as property for JenaRDFLiteral - http://bit.ly/bwxXRf
15:11 jonalv egonw: so how do I get the string I should use to check whether the factory object contributed by the EP can handle the given RDF?
15:12 jonalv egonw: the URI String...
15:16 jonalv zarah: ping egonw
15:16 zarah Consider it noted.
15:35 samuell joined #bioclipse
15:51 egonw jonalv: with some Jena magic
15:51 zarah egonw: You have new messages. Write '/msg zarah @messages' to read them.
15:51 egonw I'll take care of that
15:51 jonalv egonw: okey
15:52 jonalv egonw: I will leave some TODOs for you then.,..
15:52 egonw great
16:01 jonalv egonw: what Object should I send to the factory given by the extension point?
16:02 jonalv or ratehr what class should it be an instance of?
16:02 jonalv egonw: is it just simply the model?
16:03 egonw on conf call
16:04 jonalv oh okey
16:12 jonalv egonw: ping me when you are available
17:07 edrin left #bioclipse
17:07 egonw jonalv: ping
17:07 egonw got 10 minutes or so before dinner
17:08 jonalv egonw: okey maybe we just take it tomorrow instead?
17:08 jonalv egonw: I read the text btw
17:08 jonalv egonw: lots of code there
17:08 jonalv egonw: maybe treat the code examples as figures?
17:08 egonw yeah, I had to show something :)
17:08 egonw yeah, I might do that indeed
17:08 jonalv egonw: and give them a caption...
17:09 jonalv I think they could use captions
17:09 egonw back to your question
17:09 jonalv color coding would be sweet too :)
17:09 egonw what role does the factory have?
17:09 egonw what is it doing?
17:09 jonalv egonw: Where should the net.bioclipse.cdk.rdf plugin be?
17:09 jonalv egonw: it has two methods
17:10 jonalv boolean canHandle(String uri)
17:10 egonw jonalv: let's put these in the bioclipse.rdf repository
17:10 jonalv and IBioObject rdfToBioObject(Model model, REsource res)
17:10 egonw yes, that sounds good
17:11 egonw but as said, I like the canHandle() to be done via the extension's XML
17:11 jonalv I am not 1000 % sure about the difference between model and resource though...
17:11 jonalv egonw: I don't understand how that works
17:11 jonalv egonw: what makes you think it can be done?
17:11 egonw olass: ping
17:12 egonw olass: can you explain how to introspect an extension's XML definition?
17:12 egonw I know we need to add a field to the EPs xml schema
17:12 jonalv egonw: or hm, if the extension point contribute that string as a seperate thing
17:12 egonw but how does it go from there?
17:12 jonalv egonw: that could actually work...
17:13 jonalv egonw: I wonder if there is something like that already in Bioclispe which I can look at
17:13 egonw jonalv: I guess the QSAR feature
17:13 egonw net.bioclipse.qsar/schema
17:14 jonalv egonw: okey cool, then I know what to start with tomorrow :)
17:15 egonw I think the code to loop over the extensions is in OntologyHelper in the same plugin
17:16 egonw ah, no:
17:16 egonw QsarHelper.readProvidersAndDescriptorImplsfromEP
17:17 egonw ok, dinner time
17:17 egonw bbl
17:25 * jonalv gives up for today
18:05 samuell Hmm ... wonder why newlines are not preserved when I write a JS variable to disk ... hmm ..
18:13 * samuell wonders whether I can preserve them in some way ...
18:22 samuell Now I have to add " \n\" at the end of each line of the stuff I save in a JS variable (some prolog code), which doesn't look nice :/
19:19 sneumann joined #bioclipse
20:19 annzi joined #bioclipse
21:07 annzi joined #bioclipse
21:16 samuell joined #bioclipse
21:30 annzi joined #bioclipse

| Channels | #bioclipse index | Today | | Search | Google Search | Plain-Text | summary